Research map / Video / Video generation
Infinity-RoPE: Action-Controllable Infinite Video Generation Emerges From Autoregressive Self-Rollout
Hidir Yesiltepe, Tuna Meral, Adil Kaan Akan, Kaan Oktay, Pinar Yanardag
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| increase temporal jump index from 6 to 90 (ablation) | Infinity-RoPE RoPE Cut · RoPE Cut, 40s videos | Background Consistency: 88.98 -> 82.27 | Hurt |
| increase temporal jump index from 6 to 90 (ablation) | Infinity-RoPE RoPE Cut · RoPE Cut, 40s videos | Subject Consistency: 90.74 -> 87.69 | Hurt |
| increase temporal jump index from 21 to 45 (outside training horizon) (ablation) | Infinity-RoPE RoPE Cut · RoPE Cut, 40s videos | Background Consistency: 87.57 -> 84.48 | Hurt |
| increase temporal jump index from 6 to 90 (ablation) | Infinity-RoPE RoPE Cut · RoPE Cut, 40s videos | Temporal Smoothness: 0.98 -> 0.95 | Hurt |
| increase KV cache size (ablation) | Infinity-RoPE · MovieGenBench long videos | Imaging Quality and Dynamic Degree | Hurt |
| increase KV cache size (ablation) | Infinity-RoPE · MovieGenBench long videos | Subject and Background Consistency | No clear effect |
| replace Rolling-Forcing with Block-Relativistic RoPE on Self-Forcing | Rolling-Forcing · MovieGenBench 240s videos | VBench Overall: 0.8017 -> 0.8309 | Improved |
| apply Block-Relativistic RoPE to Self-Forcing | Self-Forcing · MovieGenBench 240s videos | VBench Overall: 0.6850 -> 0.8309 | Improved |
| apply Block-Relativistic RoPE to Self-Forcing | Self-Forcing · MovieGenBench 5s videos | VBench Overall: 0.8398 -> 0.8377 | No clear effect |
| use KV Flush instead of KV-Recache | LongLive · User study, action-controlled generation | Text Alignment: 3.19 -> 3.86 | Improved |
| apply Infinity-RoPE instead of Rolling-Forcing for 60s videos | Rolling-Forcing · User study, 60s videos | Average Likert score: 3.488 -> 3.810 | Improved |
| apply Infinity-RoPE instead of Self-Forcing for 60s videos | Self-Forcing · MovieGenBench 60s videos | VBench Overall: 0.7715 -> 0.8298 | Improved |
Later papers that compared with this method
10 later accepted papers report Infinity-RoPE as a baseline in their results tables; 10 of them report a better number on most of the results they share with it.
- Context Forcing: Consistent Autoregressive Video Generation with Long Context (ICML 2026 · better on 5 of 5 shared results)
- Anchor Forcing: Anchor Memory and Tri-Region RoPE for Interactive Streaming Video Diffusion (ECCV 2026 · better on 5 of 5 shared results)
- Mode Seeking meets Mean Seeking for Fast Long Video Generation (ICML 2026 · better on 4 of 4 shared results)
- Flex-Forcing: Towards a Unified Autoregressive and Bidirectional Video Diffusion Model (ICML 2026 · better on 3 of 3 shared results)
- What Remains in Sight? Autoregressive Video Decoding as Representation-Guided Context Rewriting (NeurIPS 2026 · better on 3 of 3 shared results)
- AdaState: Self-Evolving Anchors for Streaming Video Generation (NeurIPS 2026 · better on 3 of 3 shared results)
- VideoMLA: Low-Rank Latent KV Cache for Minute-Scale Autoregressive Video Diffusion (NeurIPS 2026 · better on 2 of 2 shared results)
- ShotStream: Streaming Multi-Shot Video Generation for Interactive Storytelling (ECCV 2026 · better on 2 of 2 shared results)
2 more not listed.
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- Rolling-Forcing — Rolling Forcing: Autoregressive Long Video Diffusion in Real Time (ICLR 2026) · better on 7 of 7 shared results
- LongLive — LongLive: Real-time Interactive Long Video Generation (ICLR 2026) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- InfinityStar: Unified Spacetime AutoRegressive Modeling for Visual Generation (NeurIPS 2025 · similarity 0.89)
- AR-Diffusion: Asynchronous Video Generation with Auto-Regressive Diffusion (CVPR 2025 · similarity 0.89)
- AdaState: Self-Evolving Anchors for Streaming Video Generation (NeurIPS 2026 · similarity 0.89)
- From Slow Bidirectional to Fast Autoregressive Video Diffusion Models (CVPR 2025 · similarity 0.88)
- Wan-Move: Motion-controllable Video Generation via Latent Trajectory Guidance (NeurIPS 2025 · similarity 0.88)
- DiTCtrl: Exploring Attention Control in Multi-Modal Diffusion Transformer for Tuning-Free Multi-Prompt Longer Video Generation (CVPR 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
